Surron—Study of Evaporation from Water-Surfaces. 147 
other two. The ratio of the evaporation in the one-foot pipe to 
that in the three-foot is about 5/4 before noon; 4/3 during the 
afternoon, and 3/2 after sunset. Evidently, then, the evaporation 
observed here is not inversely proportional to the length of the 
tube, nor anything like it. Two points are in fact doubtful 
according to these observations—one, that the air in immediate 
contact with the water is even nearly saturated with water-vapour ; 
the other, that the vapour-tension gradient in the tube is uniform. 
With regard to the second point, it isto be noted that the diffusion 
formula E = Kp'/l 
[where # is the quantity of water evaporated, 
Ff the coefficient of diffusion, 
p the vapour-tension at the temperature of the water, and 
/ the length of the tube] 
necessarily breaks down for very short tubes in air, because an 
indefinitely short tube would premise an infinitely rapid evapora- 
tion. It is, moreover, to be noted that in my experiments the 
temperature of the water is not in general the same as that of the 
surrounding air, as it perhaps mostly is in laboratory researches. 
My tentative suggestion is that the diffusion formula under 
meteorological (not laboratory) conditions only holds good in 
tubes whose length is great, measured in terms of their diameter, 
but that in other cases, provided the tubes are not too short, it 
approximates more nearly to the form 
HK = Kpi/t 
where 7 is the true vapour-tension immediately above the surface 
of the water. If this is so, then Tables 10 and 11 indicate a 
vapour-tension at the bottom of the three-foot tube twice as great 
as that at the bottom of the one-foot tube. Failing this com- 
promise, the results are better expressed by 
E> = Kp’ ft, 
although this is far from exact for the three given lengths. 
The ratios between the total quantities of evaporation from the 
four cups, i.e. the open cup and the three in the pipes, are 
3885 : 100 : 82: 68. 
A pretty good working formula for these ratios is 
E=a+bl+c +d... 
